提交 · 2e07fa9cd3bac1e28cfe3131ed86b053afb02fc9 · openeuler / Kernel

26 4月, 2007 40 次提交

[SK_BUFF]: Use offsets for skb->{mac,network,transport}_header on 64bit architectures · 2e07fa9c

由 Arnaldo Carvalho de Melo 提交于 4月 10, 2007

With this we save 8 bytes per network packet, leaving a 4 bytes hole to be used
in further shrinking work, likely with the offsetization of other pointers,
such as ->{data,tail,end}, at the cost of adds, that were minimized by the
usual practice of setting skb->{mac,nh,n}.raw to a local variable that is then
accessed multiple times in each function, it also is not more expensive than
before with regards to most of the handling of such headers, like setting one
of these headers to another (transport to network, etc), or subtracting, adding
to/from it, comparing them, etc.

Now we have this layout for sk_buff on a x86_64 machine:

[acme@mica net-2.6.22]$ pahole vmlinux sk_buff
struct sk_buff {
	struct sk_buff *       next;             /*   0   8 */
	struct sk_buff *       prev;             /*   8   8 */
	struct rb_node         rb;               /*  16  24 */
	struct sock *          sk;               /*  40   8 */
	ktime_t                tstamp;           /*  48   8 */
	struct net_device *    dev;              /*  56   8 */
	/* --- cacheline 1 boundary (64 bytes) --- */
	struct net_device *    input_dev;        /*  64   8 */
	sk_buff_data_t         transport_header; /*  72   4 */
	sk_buff_data_t         network_header;   /*  76   4 */
	sk_buff_data_t         mac_header;       /*  80   4 */

	/* XXX 4 bytes hole, try to pack */

	struct dst_entry *     dst;              /*  88   8 */
	struct sec_path *      sp;               /*  96   8 */
	char                   cb[48];           /* 104  48 */
	/* cacheline 2 boundary (128 bytes) was 24 bytes ago*/
	unsigned int           len;              /* 152   4 */
	unsigned int           data_len;         /* 156   4 */
	unsigned int           mac_len;          /* 160   4 */
	union {
		__wsum         csum;             /*       4 */
		__u32          csum_offset;      /*       4 */
	};                                       /* 164   4 */
	__u32                  priority;         /* 168   4 */
	__u8                   local_df:1;       /* 172   1 */
	__u8                   cloned:1;         /* 172   1 */
	__u8                   ip_summed:2;      /* 172   1 */
	__u8                   nohdr:1;          /* 172   1 */
	__u8                   nfctinfo:3;       /* 172   1 */
	__u8                   pkt_type:3;       /* 173   1 */
	__u8                   fclone:2;         /* 173   1 */
	__u8                   ipvs_property:1;  /* 173   1 */

	/* XXX 2 bits hole, try to pack */

	__be16                 protocol;         /* 174   2 */
	void    (*destructor)(struct sk_buff *); /* 176   8 */
	struct nf_conntrack *  nfct;             /* 184   8 */
	/* --- cacheline 3 boundary (192 bytes) --- */
	struct sk_buff *       nfct_reasm;       /* 192   8 */
	struct nf_bridge_info *nf_bridge;        /* 200   8 */
	__u16                  tc_index;         /* 208   2 */
	__u16                  tc_verd;          /* 210   2 */
	dma_cookie_t           dma_cookie;       /* 212   4 */
	__u32                  secmark;          /* 216   4 */
	__u32                  mark;             /* 220   4 */
	unsigned int           truesize;         /* 224   4 */
	atomic_t               users;            /* 228   4 */
	unsigned char *        head;             /* 232   8 */
	unsigned char *        data;             /* 240   8 */
	unsigned char *        tail;             /* 248   8 */
	/* --- cacheline 4 boundary (256 bytes) --- */
	unsigned char *        end;              /* 256   8 */
}; /* size: 264, cachelines: 5 */
   /* sum members: 260, holes: 1, sum holes: 4 */
   /* bit holes: 1, sum bit holes: 2 bits */
   /* last cacheline: 8 bytes */

On 32 bits nothing changes, and pointers continue to be used with the compiler
turning all this abstraction layer into dust. But there are some sk_buff
validation tricks that are now possible, humm... :-)
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

2e07fa9c

[SK_BUFF]: unions of just one member don't get anything done, kill them · b0e380b1

由 Arnaldo Carvalho de Melo 提交于 4月 10, 2007

Renaming skb->h to skb->transport_header, skb->nh to skb->network_header and
skb->mac to skb->mac_header, to match the names of the associated helpers
(skb[_[re]set]_{transport,network,mac}_header).
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

b0e380b1

[SK_BUFF]: Introduce skb_network_header_len · cfe1fc77

由 Arnaldo Carvalho de Melo 提交于 3月 16, 2007

For the common sequence "skb->h.raw - skb->nh.raw", similar to skb->mac_len,
that is precalculated tho, don't think we need to bloat skb with one more
member, so just use this new helper, reducing the number of non-skbuff.h
references to the layer headers even more.
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

cfe1fc77

[SK_BUFF]: Use the helpers to get the layer header pointer · bff9b61c

由 Arnaldo Carvalho de Melo 提交于 3月 16, 2007

Some more cases...
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

bff9b61c

[NET_SCHED]: Fix warning · 514bca32

由 Patrick McHardy 提交于 3月 16, 2007

net/sched/sch_api.c: In function 'psched_show':
net/sched/sch_api.c:1219: warning: format '%08x' expects type 'unsigned int', but argument 6 has type 's64'
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

514bca32

[NET_SCHED]: sch_cbq: fix watchdog scheduled too late · bb239acf

由 Patrick McHardy 提交于 3月 16, 2007

q->now is increased during dequeue and doesn't contain the current time
afterwards, resulting in a too large timeout value for the qdisc watchdog.
Use "now" instead, which still contains the current time.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

bb239acf

[NET_SCHED]: Export real timer resolution in /proc/net/psched · 4361cb17

由 Patrick McHardy 提交于 3月 16, 2007

The timer resolution exported in /proc/net/psched is used by userspace to
calculate HTB's burst values. Currently it is set to HZ, since we're now
using hrtimers, use KTIME_MONOTONIC_RES, which makes HTB use smaller burst
values.

This patch also affects libnl, which incorrectly uses this value for
the SFQ perturbation parameter, which is always in seconds, and some
routing cache values, which are in USER_HZ, so both cases are broken
anyway.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

4361cb17

[NET_SCHED]: kill jiffie conversion macros · 00c04af9

由 Patrick McHardy 提交于 3月 16, 2007

Now that all packet schedulers have been converted to hrtimers most users
of PSCHED_JIFFIE2US and PSCHED_US2JIFFIE are gone. The remaining users use
it to convert external time units to packet scheduler clock ticks, so use
PSCHED_TICKS_PER_SEC instead.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

00c04af9

[NET_SCHED]: sch_htb: use hrtimer based watchdog · fb983d45

由 Patrick McHardy 提交于 3月 16, 2007

Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

fb983d45

[NET_SCHED]: sch_cbq: use hrtimer for delay_timer · 1a13cb63

由 Patrick McHardy 提交于 3月 16, 2007

Switch delay_timer to hrtimer.

The class penalty parameter is changed to use psched ticks as units.
Since iproute never supported using this and the only existing user
(libnl) incorrectly assumes psched ticks as units anyway, this
shouldn't break anything.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

1a13cb63

[NET_SCHED]: sch_cbq: fix cbq_undelay_prio for non-active priorites · e9054a33

由 Patrick McHardy 提交于 3月 16, 2007

cbq_undelay_prio is supposed to return a time delta, but returns the
current time for non-active priorities, causing cbq_undelay to mark
the priority as active and schedule a timer for twice the current
time.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

e9054a33

[NET_SCHED]: sch_cbq: use hrtimer based watchdog · 88a99354

由 Patrick McHardy 提交于 3月 16, 2007

Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

88a99354

P
[NET_SCHED]: sch_netem: use hrtimer based watchdog · 59cb5c67
由 Patrick McHardy 提交于 3月 16, 2007
```
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
```
59cb5c67

[NET_SCHED]: sch_tbf: use hrtimer based watchdog · f7f593e3

由 Patrick McHardy 提交于 3月 16, 2007

Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

f7f593e3

P
[NET_SCHED]: sch_hfsc: use hrtimer based watchdog · ed2b229a
由 Patrick McHardy 提交于 3月 16, 2007
```
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
```
ed2b229a

[NET_SCHED]: Add hrtimer based qdisc watchdog · 4179477f

由 Patrick McHardy 提交于 3月 16, 2007

Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

4179477f

[NET_SCHED]: Use ktime as clocksource · 641b9e0e

由 Patrick McHardy 提交于 3月 16, 2007

Get rid of the manual clock source selection mess and use ktime. Also
use a scalar representation, which allows to clean up pkt_sched.h a bit
more and results in less ktime_to_ns() calls in most cases.

The PSCHED_US2JIFFIE/PSCHED_JIFFIE2US macros are implemented quite
inefficient by this patch, following patches will convert all qdiscs
to hrtimers and get rid of them entirely.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

641b9e0e

[SK_BUFF]: Some more layer header conversions · ddc7b8e3

由 Arnaldo Carvalho de Melo 提交于 3月 15, 2007

Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

ddc7b8e3

A
[KBUILD]: Unifdef headers changed by the skb layer header refactorings · 0a6114d9
由 Arnaldo Carvalho de Melo 提交于 3月 15, 2007
```
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
```
0a6114d9

[SK_BUFF]: More skb_put related skb_reset_transport_header · d10ba34b

由 Arnaldo Carvalho de Melo 提交于 3月 14, 2007

This time we have to set it to skb->tail that is not anymore equal to
skb->data, so we either add a new helper or just add the skb->tail - skb->data
offset, for now do the later.
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

d10ba34b

[IPV6]: Reset the network header in ip6_nd_hdr · 55f79cc0

由 Arnaldo Carvalho de Melo 提交于 3月 14, 2007

ip6_nd_hdr is always called immediately after a alloc_skb + skb_reserve
sequence, i.e. when skb->tail is equal to skb->data, making it correct to use
skb_reset_network_header().
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

55f79cc0

[SK_BUFF]: More skb_put related conversions to skb_reset_transport_header · eeeb0374

由 Arnaldo Carvalho de Melo 提交于 3月 14, 2007

This is similar to the skb_reset_network_header(), i.e. at the point we reset
the transport header pointer/offset skb->tail is equal to skb->data.
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

eeeb0374

[NETFILTER]: nfnetlink: parse attributes with nfattr_parse in nfnetlink_check_attribute · ac6d141d

由 Pablo Neira Ayuso 提交于 3月 14, 2007

Use nfattr_parse to parse attributes, this patch also modifies the default
behaviour since unknown attributes will be ignored instead of returning
EINVAL. This ensure backward compatibility: new libraries with new
attributes and old kernels can work.
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

ac6d141d

[NETFILTER]: ctnetlink: add support for internal tcp connection tracking flags handling · c8e2078c

由 Pablo Neira Ayuso 提交于 3月 14, 2007

This patch let userspace programs set the IP_CT_TCP_BE_LIBERAL flag to
force the pickup of established connections.
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

c8e2078c

[NETFILTER]: TCP conntrack: factorize out the PUSH flag · 5c8ce7c9

由 Willy Tarreau 提交于 3月 14, 2007

The PUSH flag is accepted with every other valid combination.
Let's get it out of the tcp_valid_flags table and reduce the
number of combinations we have to handle. This does not
significantly reduce the table size however (8 bytes).
Signed-off-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

5c8ce7c9

[NETFILTER]: TCP conntrack: accept RST|PSH as valid · 8f5bd990

由 Willy Tarreau 提交于 3月 14, 2007

This combination has been encountered on an IBM AS/400 in response
to packets sent to a closed session. There is no particular reason
to mark it invalid.
Signed-off-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

8f5bd990

[NETFILTER]: nf_conntrack: add nf_copy() to safely copy members in skb · e7ac05f3

由 Yasuyuki Kozakai 提交于 3月 14, 2007

This unifies the codes to copy netfilter related datas. Before copying,
nf_copy() puts original members in destination skb.
Signed-off-by: NYasuyuki Kozakai <yasuyuki.kozakai@toshiba.co.jp>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

e7ac05f3

[NETFILTER]: nf_conntrack: add __nf_copy() to copy members in skb · edda553c

由 Yasuyuki Kozakai 提交于 3月 14, 2007

This unifies the codes to copy netfilter related datas. Note that
__nf_copy() assumes destination skb doesn't have any netfilter
related members.
Signed-off-by: NYasuyuki Kozakai <yasuyuki.kozakai@toshiba.co.jp>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

edda553c

[NETFILTER]: nf_conntrack: use jhash2 in __hash_conntrack · 9b887909

由 Sami Farin 提交于 3月 14, 2007

Now it uses jhash, but using jhash2 would be around 3-4 times faster
(on P4).
Signed-off-by: NSami Farin <safari-netfilter@safari.iki.fi>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

9b887909

[JHASH]: Use const in jhash2 · 8e87e014

由 Patrick McHardy 提交于 3月 14, 2007

Use const to avoid forcing users to cast const data.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

8e87e014

[NETFILTER]: nfnetlink: move EXPORT_SYMBOL declarations next to the exported symbol · f4bc177f

由 Pablo Neira Ayuso 提交于 3月 14, 2007

Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

f4bc177f

[NETFILTER]: nfnetlink: remove unused includes in nfnetlink.c · 8a2e8953

由 Pablo Neira Ayuso 提交于 3月 14, 2007

Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

8a2e8953

[NETFILTER]: nfnetlink: remove unrequired check in nfnetlink_get_subsys · ac0f1d98

由 Pablo Neira Ayuso 提交于 3月 14, 2007

subsys_table is initialized to NULL, therefore just returns NULL in case
that it is not set.
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

ac0f1d98

[NETFILTER]: nfnetlink: remove duplicate checks in nfnetlink_check_attributes · d9e6d029

由 Pablo Neira Ayuso 提交于 3月 14, 2007

Remove nfnetlink_check_attributes duplicates message size and callback
id checks. nfnetlink_find_client and nfnetlink_rcv_msg already do
such checks.
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

d9e6d029

[NETFILTER]: nfnetlink: remove early debugging messages from nfnetlink · 67ca3966

由 Pablo Neira Ayuso 提交于 3月 14, 2007

Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

67ca3966

P
[NETFILTER]: nf_conntrack: uninline notifier registration functions · 010c7d6f
由 Patrick McHardy 提交于 3月 14, 2007
```
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
```
010c7d6f

[NETFILTER]: nfnetlink: use netlink_run_queue() · 73c36186

由 Patrick McHardy 提交于 3月 14, 2007

Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

73c36186

P
[NETFILTER]: nfnetlink: use mutex instead of semaphore · a3c5029c
由 Patrick McHardy 提交于 3月 14, 2007
```
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
```
a3c5029c

[NETFILTER]: nf_conntrack: simplify l4 protocol array allocation · c6a1e615

由 Patrick McHardy 提交于 3月 14, 2007

The retrying after an allocation failure is not necessary anymore
since we're holding the mutex the entire time, for the same
reason the double allocation race can't happen anymore.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

c6a1e615

[NETFILTER]: nf_conntrack: simplify protocol locking · 0661cca9

由 Patrick McHardy 提交于 3月 14, 2007

Now that we don't use nf_conntrack_lock anymore but a single mutex for
all protocol handling, no need to release and grab it again for sysctl
registration.
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

0661cca9

openeuler / Kernel 大约 1 年 前同步成功

openeuler / Kernel
大约 1 年前同步成功